Business tools
Plug predictive risk scores, route analysis, and citywide forecasting directly into your tools - over a secure API.
Point-based risk score
{
"lat": -26.2041,
"lng": 28.0473,
"date": "2026-02-07",
"horizon_days": 7,
"categories": "ALL",
"include_breakdown": true
}Click the map or run the request to generate fresh results.
Awaiting request
-One API for real-time safety intelligence
You dropped a pin on the map above and got back a risk score. That is the Safety Intelligence API - real-time and predictive location risk, available to any application, platform, or operations team through standard REST endpoints. Submit a coordinate. Score a route. Pull a city-wide forecast. All authenticated with a single API key.
Understanding the safety profile of a location currently means stitching together quarterly government reports, informal community channels, and proprietary guesswork. None of these sources talk to each other. Crime statistics are months old before they reach a spreadsheet. Community-reported incidents live in WhatsApp groups with no structure. Emergency logs stay locked inside isolated control rooms.
For insurers pricing risk, logistics operators planning routes, property platforms surfacing neighbourhood data, and city planners allocating resources - the dataset they need either does not exist or cannot be queried programmatically.
What powers it
The Safety Intelligence API draws on millions of geo-tagged safety events - community-reported incidents, verified emergency responses, access control logs, and patrol activity - aggregated and structured into a continuously learning safety graph. Raw, siloed signals become standardised, queryable risk intelligence.
This is not a static crime database. As new signals arrive, the models retrain and the intelligence sharpens. The result is a living dataset that reflects what is happening now, not what happened last quarter.
Location risk
Send any latitude and longitude. Receive a probability distribution across six standardised crime categories and a single composite risk score - in milliseconds.
Route scoring
Submit a planned route and get segment-by-segment risk scoring. Dispatch teams, logistics operators, and travel planners see high-risk stretches before anyone moves.
City forecasting
Run batch predictions across a city grid to power heatmaps, scheduled briefings, and automated alerts - with a forecast horizon of up to 40 weeks.
Insurance and underwriting
Price policies with real-time risk scores rather than annual static datasets.
Mobility and logistics
Score delivery routes, ride-hail corridors, and driver allocation zones before dispatch.
Property and real estate
Surface verified safety profiles alongside listings instead of anecdotal neighbourhood ratings.
Private security
Optimise patrol coverage, benchmark response areas, and brief teams with predictive intelligence.
City and public sector
Monitor hotspots, measure intervention impact, and allocate resources based on predicted risk.
Mapping and navigation
Embed safety layers into existing map products without building the data pipeline yourself.
How it works
Every response returns probabilities across six crime categories. They sum to 1.0 and roll up into a single composite score - so you can filter, sort, and threshold without building custom models.
Example response
{
"location": { "lat": -33.9249, "lng": 18.4241 },
"timestamp": "2026-02-07T18:00:00Z",
"categories": {
"THEFT_BURGLARY": 0.31,
"BATTERY_ASSAULT": 0.18,
"VEHICLE_DAMAGE": 0.12,
"SERIOUS_VIOLENT": 0.07,
"OTHER": 0.09,
"NO_CRIME": 0.23
},
"risk_score": 0.68
}Probabilities sum to 1.0. The composite risk score is weighted from category values.
Coverage
Johannesburg, Cape Town, Durban, London, Chicago, Los Angeles, San Francisco, New York City, Boston, and Philadelphia - with new cities added as datasets are verified.
Model performance
The prediction model uses a spatial convolutional network with temporal features, trained on verified incidents and continuously retrained as new signals arrive.
Book a demo to review endpoints, coverage, and pricing. The API is available in sandbox, partner, and enterprise tiers with bearer token authentication, tiered rate limits, and audit logging.
Prefer to self-serve?
Read the API docs
.